ARTICLE IN PRESS

Continental Shelf Research 30 (2010) 127–137

Contents lists available at ScienceDirect

Continental Shelf Research

0278-43

doi:10.1

n Corr

E-m

journal homepage: www.elsevier.com/locate/csr

Dispersion in the Yucatan coastal zone: Implications for red tide events

Cecilia Enriquez n, Ismael J Marino-Tapia, Jorge A Herrera-Silveira

Departamento de Recursos del Mar, Cinvestav, Km 6 Carretera Antigua a Progreso, Cordemex, 97310, A.P. 73, Merida, Yuc., Mexico

a r t i c l e i n f o

Article history:

Received 21 July 2008

Received in revised form

17 September 2009

Accepted 8 October 2009Available online 30 October 2009

Keywords:

Yucatan Current

Wind-driven circulation

Campeche Bank

HABs

Hydrodynamics

43/$ - see front matter & 2009 Elsevier Ltd. A

016/j.csr.2009.10.005

esponding author. Tel.: +52 999 942 94 58; f

ail address: cenriquez@mda.cinvestav.mx (C.

a b s t r a c t

The mechanisms governing dispersion processes in the northern Yucatan coast are investigated using a

barotropic numerical model of coastal circulation, which includes wind-generated and large scale

currents (i.e. Yucatan Current). This work provides the foundations for studying the dispersion of

harmful algal blooms (HABs) in the area. Modelling experiments include effects of climatic wind (from

long term monthly mean NCEP reanalysis), short term wind events (from in situ point measurements),

and Yucatan Current (YC) characteristics. Its magnitude was approximated from published reports, and

its trajectory from geostrophic current fields derived from altimeter data. These provided a range of real

and climatic conditions to study the routes in which phytoplankton blooms may travel. The 2-D model

results show that a synthetic and conservative bloom seeded in the Cabo Catoche (CC) region (where it

usually grows), moves along the coast to the west up to San Felipe (SF), where it can either move

offshore, or carry on travelling westwards. The transport to the west up to SF is greatly influenced by

the trajectory, intensity and proximity of the YC jet to the peninsula, which enhances the westward

circulation in the Yucatan Shelf. Numerical experiments show that patch dispersion is consistently to

the west even under the influence of northerly winds. When the YC flows westward towards the

Campeche Bank, momentum transfer caused by the YC jet dominates the dispersion processes over

wind stress. On the other hand, when it flows closer to Cuba, the local processes (i.e. wind and

bathymetry) become dominant. Coastal orientation and the Coriolis force may be responsible for

driving the patch offshore at SF if external forcing decreases.

& 2009 Elsevier Ltd. All rights reserved.

1. Introduction

The coastal sea of the northern Yucatan Peninsula (Fig. 1) has awide and shallow continental shelf (up to 245 km wide with anearly monotonic 1/1000 slope). It is located between theCaribbean Sea and the Gulf of Mexico, two ecosystemscommunicated through the Yucatan Channel, which is 196 kmwide and reaches 2000 m depth. The YC flows through thischannel, carrying with it different water masses. This current cangenerate a dynamic upwelling pushing cold and nutrient-richwater uphill across the steep continental slope, reaching theYucatan Shelf where it is dispersed at the bottom (Cochrane,1969; Merino, 1992; Merino 1997; Ruiz-Renteria, 1979; Sahlet al., 1997). This upwelling provides conditions for developmentof algal blooms and the resulting food web enhancement thatsupports species such as the whale shark (Rhincodon typus) in theCabo Catoche (CC) region (Fig. 1). In addition to the upwellingprocesses, nutrients may be supplied by groundwater discharges.Continental water in the region drains to the sea as submarinegroundwater discharges (SGD), which happen in several locations

ll rights reserved.

ax: +52 999 981 23 34.

Enriquez).

such as Dzilam Bravo (DB). There is some evidence showingthat algal blooms are fed initially by the upwelling at CC andenhanced by SGD as they travel to the west near the populatedareas of the northern coast of Yucatan (Alvarez-Gongora, 2009;Herrera-Silvera et al., 2004).

Nonetheless, under certain circumstances algal blooms devel-op harmful characteristics (HABs) creating havoc on the localenvironment, fisheries and tourist industry (Alvarez-Gongora andHerrera-Silveira, 2006; Herrera-Silvera et al.,2004). In Yucatan,phytoplankton blooms have been recorded since 1948. The mostrecent and harmful events happened in 2001, 2003 and 2008.Based on the damage caused by these events, research andmonitoring programs were implemented to determine thehydrological conditions and oceanographic processes related withthe frequency, spatial distribution and species composition ofHABs in order to understand their behaviour and minimize thenegative ecological and economic impacts. The ultimate goal is tohave a forecast system to aid an adequate management of red tidethreats.

So far, red tide studies have focused on hydrological char-acterization, taxonomy of HAB species, water quality andecological impacts (Alvarez-Gongora and Herrera-Silveira, 2006;Hernandez-Becerril et al., 2007). However, hydrodynamics andthe resulting dispersion are poorly understood in the Yucatan

ARTICLE IN PRESS

Fig. 2. Model grid, showing the open boundaries. The figure also shows the red

tide discharge location, marked with a black dot and the observation points A, B, C

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137128

Shelf, which is the north-eastern region of the so-called CampecheBank. The information available in the literature regarding thehydrodynamics in this region is scarce. Capurro-Filograsso andReid (1972) provided the first collection of oceanographic studiesfrom the Gulf of Mexico; most of these studies were large scaleoceanographic surveys where the Yucatan Shelf was coveredpartially or not included. Merino (1992; 1997) conducted a seriesof field campaigns measuring thermohaline distribution in theYucatan Shelf, particularly in the Cabo Catoche region, to studythe upwelling process. The currents have not been studied indetail in the Yucatan Shelf but it is commonly known by the localsthat the dominant current in the coastal region is to the west. Thewestward current has been mentioned in other studies aroundthe region (Merino, 1997; Monreal, et al., 1992). More recently,Zavala-Hidalgo et al. (2003) and later Morey et al. (2005) studiedthe circulation on continental shelves of the Gulf of Mexico usingnumerical modelling forced with climatic winds. Both confirm theknown westward circulation, which coincides with a dominantwesterly wind on the region throughout the year. Morey et al.(2005) suggest that over wide shelves, such as the Yucatan Shelf,sub tidal variability of circulation is dominantly forced by localsurface momentum fluxes (wind stress). Zavala-Hidalgo et al.(2003) presented evidence that this applies for most of the GOMcontinental shelves. Using 7 years of model data, they found highcorrelation (values r40.5) between monthly mean surfacecurrents and the along-coast wind stress component at tenlocations around the GOM, demonstrating the importance of windstress as the main force in coastal circulation. The exception tothis behaviour was their eastern-most site, Puerto Progreso,located on the Yucatan Shelf. At this site the along-coast windstress and monthly mean surface currents share sign (westward)but its correlation is the lowest (0.13). This is an indication thatover the Yucatan Shelf, processes other than wind stress musthave significant influence on the westward current behaviour.

This study gives a first step (including only hydrodynamics) toexplore the way in which different forces contribute to the localcurrent field and ultimately to the dispersion of phytoplanktonblooms. This is achieved by testing the sensitivity of the currentsto two processes: the local wind blowing over the wide andshallow continental shelf and the characteristics of the YC, mainlyits position, trajectory and intensity, using the DELFT3-Dnumerical model.

The details of the model set up and parameters used in thedifferent runs are presented in Section 2. Section 3 describes thewind and YC variability, which are the forces used to drive thenumerical experiments. Section 4 gives the analysis of red tidedispersion and includes discussions on model validation and

Fig. 1. Map of the Yucatan Shelf (Campeche Bank) showing the bottom

topography used in this study. It also shows the location of Progreso (1), Telchac

(2), Dzilam Bravo (3), San Felipe (4), Rio Lagartos (5) and Cabo Catoche (6) at the

coast.

comparison with observational data. The paper ends withconclusions in Section 5.

2. Model set up

The 2-D version of the flow module of Delft3-D numericalmodel developed by WL/Delft Hydraulics [http://delftsoftware.wldelft.nl/index.php] is used to estimate the circulation and theresulting dispersion of a synthetic and conservative (no growth/decay included) red tide event introduced in the Cabo Catoche(CC) region, where this event typically grows.

The model was implemented for the Yucatan continental shelf(northern Campeche Bank) in a fan-shaped grid with radial lengthsimilar to the Yucatan Channel length. The curvilinear grid of186�92 grid points ranges from 211 to 241 latitude and 2671 to2761 longitude (Fig. 2) with a spatial resolution, which goes fromapproximately 3 km near the coast to 5.5 km at the offshoreboundary. The model is barotropic and does not include densitydistribution at this stage, but 4 vertical sigma layers were set toallow vertical variable current forcing at the eastern boundary.After a series of sensitivity tests, the adequate time step forachieving stable results was 10 min. The bed roughness wasestimated with the Chezy formula and a cyclic advection schemewas used for momentum and transport. The values for horizontaleddy diffusivity and eddy viscosity were 10 and 1 m2/s,respectively. Bottom topography was extracted from theETOPO5 database (NOAA/AOML 1988) and complemented withhigher resolution bathymetric data with along-coast resolution of

and D located near the coast.

Table 1Detail of the model scenarios to study red tide dispersion.

ARTICLE IN PRESS

Fig. 3. Geostrophic current field obtained from sea height anomaly derived from JASON-1, TOPEX, ERS-2, ENVISAT and GFO altimeters, processed by the NRL site at the

Stennis Space Center showing different shapes in the YC path. Current field for (a) January 1, 2003, (b) March 5, 2003 and (c) May 6, 2006. Maps show the 100 and 300 m

isobaths.

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137 129

5 km measured from the coast up to the 10 m isobath (�10 kmoffshore).

The model has 3 lateral open boundaries (Fig. 2): the YucatanChannel (east), the Gulf of Mexico (north) and the Campecheboundary (west). The Yucatan Channel boundary was forced withconstant currents distributed in three sections to modulate theintensity distribution of the YC using the results of Abascal et al.(2003) and Badan et al. (2005) as guidance. The first section (YC1)starts with 0.3 m/s close to CC in the Yucatan Peninsula and endswith 0.6 m/s at about 100 km offshore. The main jet of the YCflows through the middle section (YC2), which was prescribedwith a current of 1.28 to 1.2 m/s from 100 to 185 km offshore. Theeastern section (YC3) has 0.6 m/s in the central part of theYucatan Channel and finishes with 0.3 m/s near Cabo San Antonioin Cuba. All the sections were set with an exponential decay in thevertical due to the inclusion of density invariant sigma layers.

The Gulf of Mexico and Campeche boundaries were openedwithout prescribing any forces but allowing the transit ofoutgoing waves, such as short wave disturbances, without beingreflected back into the computational domain (Verboom and Slob,1984). The Gulf of Mexico boundary is used as the tool to regulatethe path which the YC will take on its journey towards the north.This is achieved by leaving the boundary completely opened fromCampeche to Cuba or partially opened from Campeche to the

north of the Cabo Catoche area (Fig. 2, see Section 3 for a detailedexplanation).

At the surface, the model is forced with a series of windprescriptions chosen from the analysis of wind data from 2sources: (1) In situ measurements from the National MeteorologyService of the National Water Council (CNA) coastal station in RioLagartos (RL) [http://smn.cna.gob.mx/] at geographical coordi-nates 21.571 latitude and �88.161 longitude (Fig. 1). The data arehourly measurements during the year 2003. (2) The second database consists of 10 years (1997 to 2007) of reanalysis wind datafrom the National Center for Environmental Prediction (NCEP),which has 1/4 degree spatial resolution and a 6 h temporalinterval [http://www.ncep.noaa.gov/]. In all numerical experi-ments a synthetic red tide was introduced in the Cabo Catocheregion to estimate the possible routes, which the patch may takeunder different environmental conditions. The red tide in themodel is conservative and was introduced at the start of each runas a discharge of 100 kg/m3 during 30 min in one mesh pointapproximately 10 km offshore (Fig. 2). These arbitrary parametersaimed to achieve noticeable shapes by dispersion under thedifferent numerical experiments. Based on the YC patterns andthe results of wind data analysis, ten numerical experiments werechosen to summarize the results of this contribution (Table 1).Additionally the model was tested experimentally during an

ARTICLE IN PRESS

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137130

intense red tide event in the summer of 2008. The initial conditionof the red tide patch for this validation run (Run 11, Table 1) wasdetermined from the combination of chlorophyll satellite images(AQUA-MODIS data courtesy of USDOC/NOAA/NESDIS CoastWatch),direct visualization from local flights in the region andinformation received from local fishermen. The intensity of theYC was maintained as explained before, and the trajectory of thecurrent jet was emulated from the geostrophic current field ob-served in the satellite products [http://www.aoml.noaa.gov/phod/dataphod/work/trinanes/INTERFACE/index.html] on the initialday by partially closing the GOM boundary to force the currentjet to the northwest. The model was forced with the windprediction for Telchac, Yucatan obtained online from theWindGuru web site (http://www.windguru.cz/). Details of boththe initial condition of the red tide patch and the forcing of thisparticular model run are presented in Section 4.

Fig. 5. Monthly wind data averaged over 10 years (1997 to 2007) and over this

study’s domain (longitude 268.125 to 275.625 and latitude 21.90 to 23.80 N). It is

reanalysis wind data from the National Centers for Environmental Prediction (NCEP).

3. The Yucatan current and wind force

As mentioned before, the YC and the wind force are the twovariables used in this work to explore the dispersion of red tideblooms in the Yucatan Shelf. The reason for choosing these twovariables as a starting point is described in this section.

The YC is a very energetic deep current flowing adjacent to theYucatan Shelf. It affects the circulation over the shelf both, bydirectly introducing deep nutrient-rich water from the Caribbean,which generate upwelling, as proved by Cochrane (1969), Ruiz-Renteria (1979), and Merino (1992, 1997), and also (the topic ofthis paper) by introducing momentum to the local waters,especially when the YC meanders towards the west in thedirection of the Campeche Bank. The mechanisms by whichmomentum transfer occurs need further investigation, butsuggestions on the likely dominant processes are included inSection 4. The path of the YC jet variability could be associatedwith the passing of eddies through the Yucatan Channel (Badanet al., 2005). The present contribution suggests that depending onthe different attributes of the main jet, e.g. closeness to the

Fig. 4. Modelled current patterns resulting with (a) free flow into the Gulf of Mexico

boundary and (b) partially closed boundary restringing the direct flow to the north.

Yucatan Peninsula, path towards the Gulf of Mexico, etc., the YCcan exert a major influence in the currents on the Yucatan Shelf.Qualitative analysis of the geostrophic current field obtainedfrom sea height anomaly derived from JASON-1, TOPEX, ERS-2,ENVISAT and GFO altimeters, processed by the NRL site at theStennis Space Center [http://www.aoml.noaa.gov/phod/dataphod/work/trinanes/INTERFACE/index.html], shows that the YCtakes different shapes while travelling towards the north beforebecoming the well-known loop current (e.g.Capurro-Filograssoand Reid, 1972; Candela et al., 2002; Leben, 2005). It may go to thenorthwest following the isobaths on the Yucatan continentalmargin (Fig. 3a), it may travel to the north—northeast with animportant part of the jet flowing along the northern coast of Cuba(Fig. 3b) and it may take more meandered paths, for example,turning to the west adapting its path to the presence of largegyres located in the north of Cuba (Fig. 3c). These gyres may act astrue boundaries, forcing the currents to the west. To simulate thepresence (absence) of this gyres and their effect on the path of theYC, the Gulf of Mexico boundary (GOM) is partially (completely)opened (Fig. 2), restricting the flow of the main current jet, whichis forced towards the west (Fig. 4).

Ten years of NCEP wind data analysis show that the wind inthe Yucatan coastal area comes from a limited range of directions(northwest to southeast). Its seasonal variability consists ofhaving strong, short-lived wind events from the north andnortheast from October to February (Fig. 5). During that time,the north wind events alternate with south-easterlies andeasterlies, which happen all year round. The use of climatic dataas main forcing in hydrodynamic numerical models may not beenough to study short-term mesoscale events and the inputof high temporal resolution in the wind stress is extremelyvaluable to properly understand the behaviour of a system undersuch conditions (Stanev et al., 1995; Staneva and Stanev, 1998;Drakopoulos and Lascaratos, 1999). Therefore this study includeshigh frequency locally measured data to assess the importance ofwind stress on circulation during individual events. We mustmention that extremely high wind events may occur in the areaduring the hurricane season where the wind velocity mayincrease by an order of magnitude. The effect of these meteoro-logical phenomena in the dispersion is not contemplated forthis study.

The numerical experiments use a number of wind casesdetailed on Table 1. Experiments 1 to 6 include the differentwind directions by using climatic data (NCEP) of selected months.

ARTICLE IN PRESS

Fig. 7. 3-D histogram showing percentage of the wind occurrences recorded in Rio

Lagartos Station for January, 2003.

Fig. 8. 3-D histogram showing percentage of the wind occurrences recorded in Rio

Lagartos Station for June, 2003.

Fig. 6. Dispersion of the phytoplankton bloom forced with climatic monthly mean wind data from different directions (top-NE, middle-E, and bottom-SE) and with the GM

boundary totally opened (left) and partially closed (right).

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137 131

ARTICLE IN PRESS

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137132

Experiments 7 to 10 include the temporal variability of the windby using data from a meteorological station at Rio Lagartos (RL),Yucatan. All of them are tested allowing free flow to the north(totally opened GOM boundary) and with a restricted flow to thenorth (partially opened GOM boundary).

4. Dispersion of a phytoplankton bloom

The climatic wind data selected to embrace the variety ofconditions are those of January, June and July, providing windfrom the northeast, southeast and east, respectively (see Fig. 5).

There is a considerable contrast in the dispersion obtainedwith the northern boundary (GOM) completely opened(Fig. 6—left panels), and when it is partially closed (Fig. 6—rightpanels). In the first case, the model allows a free flow of thecurrent jet to the north and northeast and the dispersion in theYucatan Shelf is modulated mainly by the wind direction withlittle effect of the YC jet. In all cases (Fig. 6a, c and e) the bloom istransported to the west reaching no more than 200 km after onemonth simulations. The wind from the northeast and from thesoutheast result in more elongated shapes of the phytoplanktonpatch aligned parallel to the coast and attached to it (Fig. 6a ande). The wind coming from the east results in a more rounded

Fig. 9. Dispersion of the phytoplankton bloom after 15 days including the temporal vari

boundary totally opened (left) and partially closed (right).

shape with an offshore trend (Fig. 6c). On the other hand, thedispersion of the phytoplankton patch is very sensitive to the pathof the YC. For the partially closed GOM boundary the flow isrestricted and the current meanders to the west-northwest. It isclear that this meandering has big influence on the currents overthe continental shelf, as the dispersion of the patch is determinedby this process (Fig. 6—right panels).

It is evident that climatic data smooths both magnitude anddirection of the wind and therefore the effect on the dispersioncould be underestimated. Hence, the months of January andJune of the year 2003 were selected to include real wind events.The wind data recorded in Rio Lagartos station during Januaryis dominated by northerlies from a range between 315 and90 degrees (Fig. 7). There were also wind events from thesoutheast quadrant and very few events from the southwestquadrant. The wind coming from the north is in general moreenergetic than the rest of the wind events with velocities from2 to 10 m/s.

During June the wind was arriving from the east and southeastin a range between 50 and 180 degrees (Fig. 8) with no windevents from other directions. Wind arriving in the south-easternquadrant dominated the month with velocities ranging between 2and 9 m/s. North-easterlies and eastern wind events were fewerbut in general presented higher velocities (5–9 m/s).

ability of the wind during January (top) and June (bottom) of 2003 and with the GM

ARTICLE IN PRESS

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137 133

The effect of the wind variability in the patch dispersion isappreciated in Figs. 9 and 10. Again, when the YC flows with animportant west component, it has major influence in the bloomdispersion even though strong wind events also have directimpact on it. In all cases, the time varying wind drives thephytoplankton bloom further to the west after 30 days ofsimulation (Fig. 10) as compared to the previous experimentswith climatic winds.

On the other hand, the dispersion results produced with time-varying wind of July and June have important differences as well.For example, when the YC flows to the NE (GOM boundary totallyopened) the phytoplankton bloom persisted 4 times longer in thevicinity of the developing area (Observation point A-Fig. 2) withthe wind of January (Fig. 11a) than with the wind of June(Fig. 11c). At the end of January, the patch barely reaches the areaof Rio Lagartos (Observation point B-Fig. 2) whilst it travels100 km more to the west during June. When the YC flows tothe NW (GOM boundary partially opened) (Fig. 11b, d), thebloom patch travels all the way to Dzilam Bravo (in January) andTelchac (in June). The difference induced by the wind in thesecases is less important both in terms of the concentration andpersistence of the phytoplankton bloom at the differentobservation points.

Fig. 10. Dispersion of the phytoplankton bloom after 30 days including the temporal va

GM boundary totally opened (left) and partially closed (right).

The results of this study show that the trajectory of the YC hasa major influence on the hydrodynamics of the Yucatan Shelf.When the YC flows to the NE, patch dispersion is dominated bywind stress, which drives the patch consistently to the west. Thisbehaviour agrees with the suggested high correlation between thewind and currents in the continental shelves of the GOM reportedby other authors (i.e. Zavala-Hidalgo et al., 2003; Morey et al.,2005). However, this study shows that this correlation is weakwhen the YC flows towards the NW as the momentum transferredto the shelf is large and patch dispersion moves faster to the westadding to the effects of wind stress. This result explains the lowcorrelation between the along-coast wind stress component andthe monthly mean surface currents found by Zavala-Hidalgo et al.(2003) at Progreso, Yucatan. At this site both, along-coast windstress and monthly mean surface currents share sign (westwardflow) but its correlation is very low (0.13). This is an indicationthat currents over the Yucatan Shelf are modulated andinfluenced not only by the wind, but also in an important way,by the YC. The mechanisms by which this occurs deserves properestimation of the terms in the momentum balance equation, butlikely processes include: (i) due to its magnitude, the YC generatesa sharp pressure gradient on the eastern boundary of the YucatanShelf, the forces resulting from the sea level slope may contribute

riability of the wind during January (top) and June (bottom) of 2003 and with the

ARTICLE IN PRESS

Fig. 11. Time series of concentration (kg/m3) of the synthetic phytoplankton bloom at observation points along the coastal zone (Fig. 2) during a 1 month run for

experiments 7(a), 8(b), 9(c) and 10(d).

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137134

on driving the currents towards the west despite the winddirection. It has been shown in other continental shelves (Li andWeisberg, 1999a, b) that near shore the momentum balance isbetween the wind stress and the pressure gradient with theCoriolis term playing a secondary role. Preliminary analysis of ourresults shows that this also occurs in the Yucatan eastern shelf. (ii)the YC introduces important volumes of water into the YucatanShelf by upwelling processes, while thermohaline effects primar-ily give rise to motion with a vertical component, horizontalmotions to the west may result from continuity to allow outflow,as the YC acts as a physical boundary restricting the flow to theeast.

The opportunity to validate the hydrodynamic model cameduring the summer of 2008 when an intense red tide bloom waspresent in the Yucatan Shelf and the model could be testedexperimentally to aid the local government in the decisionmaking process. The model was initiated on July 11, when thered tide patch had already a large coverage in the coastal region.In this case the initial condition was set with the same initialconcentration/volume characteristics than the rest of the experi-

ments but over multiple different mesh points along the region.The geostrophic current field for that day (Fig. 12) showed thatthe YC was flowing to the northwest approximately along themodel’s GOM boundary. Hence, the GOM boundary was partiallyclosed from Cuba up to longitude 871 W to force the current jet toa similar path to the one observed in the geostrophic field.

The initial red tide condition was approximated based ondirect visualization of the bloom from a local flight, informationreceived from fishermen, and with the remote sensing chlorophyllconcentration image of the day (Fig. 13a). At that time, the coastalzone had two distinct large algal bloom patches. The initial redtide patches in the model at the beginning of the simulation areshown in Fig. 13c. The western region had a wide patch, whichdeveloped in the eastern corner and was dispersed to the westattached to the coast. This first bloom severely attacked theYucatan coastal zone during June and July with damaging effectson the environment and the local economy. When the secondbloom was detected as a potential threat to the already weakenedcoast, the importance of trying to predict the dispersion of thebloom was clear. Being aware of the limitations of the model

ARTICLE IN PRESS

Fig. 12. Geostrophic current field for July 11, 2008 obtained from sea height anomaly derived from JASON-1, TOPEX, ERS-2, ENVISAT and GFO altimeters, processed by the

NRL site at the Stennis Space Center.

Fig. 13. Chlorophyll concentration images of the Yucatan Shelf for (a) July 11, 2008 and (b) July 25, 2008. Both are AQUA-MODIS data courtesy of USDOC/NOAA/NESDIS

CoastWatch. The bottom panels show the modelled red tide patches (c) at the beginning of the run (July 11, 2008) and (d) at the end of the run (July 25, 2008).

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137 135

ARTICLE IN PRESS

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137136

(conservative tracer with no grow/decay behaviour and effects oflimited processes for the hydrodynamics), the model was used asa tool to predict the path of the bloom in order to provide the localauthorities with additional information to estimate the ecologicimpact of these phenomena. The result of the 15 day simulation(Fig. 13d) showed that the western bloom would continue itsdispersion to the west and southwest attached to the coast downto the Campeche region while the eastern patch was likely todetach from the coast, disperse, and move offshore. The cloudcoverage did not help to keep track of the dispersion during thefollowing days, but the local flights and official informationconfirmed what the model predicted. The chlorophyll image ofthe next cloud free day (25 July 2008) showed that the easternhigh concentration patch moved offshore and dispersedconsiderably (Fig. 13b), as predicted by the model. Similarly, thealgal bloom seeded on the western end of the peninsula travelledattached to the coast to the southwest, with propagation of‘‘fingers’’ towards the NW. The wind predicted (used for themodel) coincided with the real moderate easterly wind, whichblew in the region at that time. Considering the limitations of themodel, the above results show its capability to reproduceobserved behaviour to a satisfactory standard that served localauthorities as guidance in the decision making process.

5. Conclusions

Our study concludes that the currents over the Yucatan Shelfare modulated and influenced not only by the wind, as it wasbelieved, but also in an important way, by the momentumreceived from the YC.

The YC may travel to the northwest, north or northeast on itsway through the Yucatan Channel influenced by the character-istics of the general circulation in the eastern Gulf of Mexico andparticularly influenced by the large scale gyres located near theYucatan Channel area. The path and intensity of the YC have directinfluence in the currents over the Yucatan Shelf and consequentlyon the dispersion processes. Particularly important is thetrajectory of the YC jet and its closeness to the Yucatan Peninsula.When the YC is blocked by eddies on the north of Cuba, the YCmoves to the west enhancing the circulation over the YucatanShelf and promoting a rapid dispersion of a red tide patch. In thesecases, the wind stress only modulates the dispersion behaviourand the effect is fairly small compared to the influence of the YC.For example, wind flowing from the northwest would limit thedispersion processes whilst eastern winds will enhance them.

On the other hand, if the YC travels unblocked to the northeasttowards the strait of Florida, it has less effect on the Yucatan Shelf,and the local wind and bathymetric characteristics play moreimportant roles on the patch dispersion. Our work shows thatunder strong (short term) wind events and/or when the YC flowsaway from the Yucatan Peninsula, wind stress can be equally ormore important than the influence from the adjacent YC. In bothcases, the wind has an important contribution to the dispersion ofa red tide bloom in relation to its path, velocity and shape, andplays a determining roll in pushing the red tide patches towardsthe shore.

Results show that a patch seeded in Cabo Catoche travelsconsistently to the west, close to the coast, up to San Feliperegion and at this point it may take different routes determinedby the characteristics of the forcing. If the westward currents aredominant over the continental shelf, (e.g. the YC meanderstowards the west-northwest) then the patch may continuetravelling to the west along the coast. If the YC flows to thenortheast, or away from the Yucatan Coastal region, then it islikely that the patch may travel offshore after reaching San Felipe.

Model validation was possible for a red tide event occurred inYucatan during the summer of 2008 using web-based windpredictions and the position and path of the YC obtained form thegeostrophic field. The YC was flowing close to Cuba towards thenorthwest more than 100 km away from the Yucatan Peninsula. Amoderate wind was expected to arrive from the east. Under theseconditions (low influence of the YC together with low wind stress)the eastern patch had an offshore tendency suggesting a moreinertial behaviour, and the western patch migrated in bothdirections, offshore and to the SW along the shore.

This is the first step in studying the red tide dispersion inYucatan, future work will include a number of efforts toincorporate the effect of additional forces contributing to thelocal currents, three-dimensional thermohaline dynamics, sub-marine groundwater discharges, water exchange between the seaand coastal lagoons, and ultimately, growth and decaying rates forthe red tide, among others.

Acknowledgements

This research was supported by the CONACYT-FOMIX,YUCATAN grant No. 21336. The authors acknowledge CONANPfor support received. Wind data from the Rio Lagartos meteostation was obtained from the National Meteorology Service ofthe National Water Council (CNA). SeaWiFS data are used here inaccordance with the SeaWiFS Research Data Use Terms andConditions Agreement of the SeaWiFS project. Special thanks toDr. Gilberto Geronimo for the stimulating discussions on seasurface slope and circulation on the Yucatan Shelf.

References

Abascal, A.J., Sheinbaum, J., Candela, J., Ochoa, J., Badan, A., 2003. Analysis of flowvariability in the Yucatan Channel. J. Geophys. Res. Oceans 108 (C12), 18.

Alvarez-Gongora, C., Herrera-Silveira, J.A., 2006. Variations of phytoplanktoncommunity structure related to water quality trends in a tropical karsticcoastal zone. Mar. Pollut. Bull. 52 (1), 48–60.

Alvarez-Gongora, C. 2009. Spatial and temporal patterns of the harmful algalblooms in the north coast of Yucatan (SE, Mexico). PhD Thesis. Marine Science.CINVESTAV-IPN, Merida.

Badan, A., Candela, J., Sheinbaum, J., Ochoa, J., 2005. Upper-layer circulation in theapproaches to Yucatan Channel. In: Sturges, W., Lugo-Fernandez, A. (Eds.),Circulation in the Gulf of Mexico. Observations and Models. AGU Books Board,Geophysical monograph 161, Washington, DC, pp. 57–69.

Candela, J., Sheinbaum, J., Ochoa, J., Badan, A., Leben, R., 2002. The potentialvorticity flux through the Yucatan Channel and the Loop Current in the Gulf ofMexico. Geophys. Res. Lett. 29 (22), 4.

Capurro-Filograsso, L., Reid, J.L., 1972. Contribution on the Physical Oceanographyof the Gulf of Mexico. Gulf Publishing Co., Houston, TX.

Cochrane, J.D., 1969. Water and circulation on Campeche Bank in May. Papers inDedication to Professor Michitaka Uda., Special Bulletin of the Japanese Societyof Fisheries Oceanography, 123–129.

Drakopoulos, P.G., Lascaratos, A., 1999. Modelling the Mediterranean Sea:Climatological forcing. J. Mar. Syst. 20 (1–4), 157–173.

Ruiz-Renteria F. 1979. Upwelling north of the Yucatan Peninsula. MSc. Thesis.Texas A&M University. 86 pp.

Hernandez-Becerril, D.U., Alonso-Rodriquez, R., Alvarez-Gongora, C., Baron-Campis, S., Ceballos-Corona, G., Herrera-Silveira, J., Del Castillo, M., Juarez-Ruiz, N., Merino-Virgilio, F., Morales-Blake, A., Ochoa, J., Orellana-Cepeda, E.,Ramirez-Camarena, C., Rodriguez-Salvadoro, R., 2007. Toxic and harmfulmarine phytoplankton and microalgae (HABs) in Mexican Coasts. J. Environ.Sci. Health Part A—Toxic/Hazardous Subst. Environ. Eng. 42 (10), 1349–1363.

Herrera-Silvera, J.A., Comin, F., Aranda-Cirerol, N., Troccoli, L., Capurro, L., 2004.Coastal water quality assessment in the Yucatan Peninsula: managementimplications. Ocean Coastal Manage. 47 (11–12), 625–639.

Leben, R., 2005. Altimeter-Derived Loop Current Metrics, in Circulation in the Gulfof Mexico. In: Sturges, W., Lugo-Fernandez, A. (Eds.), Circulation in the Gulf ofMexico. Observations and Models. AGU Books Board, Geophysical monograph161, Washington, DC, pp. 181–201.

Li, Z., Weisberg, R.H., 1999a. West Florida Shelf response to upwelling favorablewind forcing: kinematics. J. Geophys. Res. 104, 13507–13527.

Li, Z., Weisberg, R.H., 1999b. West Florida Shelf response to upwelling favorablewind forcing, 2, Dynamics. J. Geophys. Res. 104, 23427–23442.

ARTICLE IN PRESS

C. Enriquez et al. / Continental Shelf Research 30 (2010) 127–137 137

Merino, M. 1992. Afloramiento en la plataforma de Yucatan. Estructura yfertilizacion., PhD. Thesis UNAM, Mexico, D. F. 255 pp.

Merino, M., 1997. Upwelling on the Yucatan Shelf: hydrographic evidence. J. Mar.Syst. 13 (1–4), 101–121.

Monreal, M.A., Salas, D.A., Padilla, A.R., Alatorre, M.A., 1992. Hydrography andestimation of density currents in the southern part of the Bay of Campeche,Mexico. Cienc. Mar. 18, 115–133.

Morey, S.L., Zavala-Hidalgo, J., O�Brien, J.J., 2005. The Seasonal Variability ofContinental Shelf Circulation in the Northern and Western Gulf of Mexico froma High-Resolution Numerical Model. In: Sturges, W., Lugo-Fernandez, A. (Eds.),Circulation in the Gulf of Mexico. Observations and Models. AGU Books Board,Geophysical monograph 161, Washington, DC, pp. 203–218.

NOAA/AOML,1988.Geostrophic current field processed by the NRL site at theStennis Space Center obtained from NOAA, Data Announcement 88-MGG-02,Digital relief of the Surface of the Earth. NOAA, National Geophysical Data

Center, Boulder, Colorado, on line dataset, /http://www.aoml.noaa.gov/phod/index.phpS.

Sahl, L.E., Wiesenburg, D., Al, Merrell, W.J., 1997. Interactions of mesoscale featureswith Texas shelf and slope waters. Cont. Shelf Res. 17 (2), 117–136.

Stanev, E.V., Roussenov, V., Rachev, N., Staneva, J., 1995. Sea response toatmospheric variability - model study for the Black-Sea. J. Mar. Systems 6(3), 241–267.

Staneva, J.V., Stanev, E.V., 1998. Oceanic response to atmospheric forcing derivedfrom different climatic data sets. Intercomparison study for the Black Sea.Oceanologica Acta 21 (3), 393–417.

Verboom, G.K., Slob, A., 1984. Weakly-reflective boundary-conditions for 2-dimensional shallow-water flow problems. Adv. Water Resour. 7 (4), 192–197.

Zavala-Hidalgo, J., Morey, S.L., O’Brien, J.J., 2003. Seasonal circulation on thewestern shelf of the Gulf of Mexico using a high-resolution numerical model.J. Geophys. Res. Oceans 108 (C12).